recipe-save-email-to-doc
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to execute commands using the
gwsCLI tool. These commands are legitimate interactions with Google Workspace services as part of the skill's intended functionality.- [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes untrusted data from Gmail messages. - Ingestion points: Step 2 in
SKILL.mdfetches message content from Gmail. - Boundary markers: Absent; the content is written directly to a Google Doc without delimiters or instructions to ignore embedded commands.
- Capability inventory: The skill uses
gws docsto create and write content to documents. - Sanitization: No sanitization of the email body is performed before it is written to the destination document.- [SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration were detected. All external tool references align with the official vendor (googleworkspace).
Audit Metadata